The present invention relates to a method and device for processing a stereo signal obtained from an encoder, which encoder encodes an N-channel audio signal into left and right signals and spatial parameters. The invention also relates to an encoder apparatus comprising such an encoder and such a device.
The present invention also relates to a method and device for processing a stereo signal obtained by such a method and such a device for processing a stereo signal obtained from an encoder. The invention also relates to a decoder apparatus comprising such a device for processing a stereo signal.
The present invention also relates to an audio system comprising such an encoder apparatus and such a decoder apparatus.
For a long time, stereo reproduction of music, for example in home environment has been prevailing. During the 1970's, some experiments were done with four channel reproduction of home music equipment.
In larger halls, such as film theatres, multi-channel reproduction of sound has been present for a long time. Dolby Digital® and other systems were developed for providing realistic and impressive sound reproduction in a large hall.
Such multi-channel systems have been introduced in the home theatre and are gaining large interest. Thus, systems having five full-range channels and one part-range channel or low-frequency effects (LFE) channel, so called 5.1 systems, are today common on the market. Other systems also exist, such as 2.1, 4.1, 7.1 and even 8.1.
With the introduction of SACD and DVD, multi-channel audio reproduction is gaining further interest. Many consumers already have the possibility of multi-channel playback in their homes, and multi-channel source material is becoming popular.
Because of increased popularity of multi-channel material, efficient coding of multi-channel material is becoming more important, which is also recognized by standardization bodies such as MPEG.
Previously known encoders often do not apply efficient methods to encode multi-channel audio. The input channels may be basically encoded individually (possibly after matrixing), thus requiring a high bit rate due to the large number of channels.
However, a multi-channel audio encoder may generate a 2-channel down-mix which is compatible with 2-channel reproduction systems, while still enabling high-quality multi-channel reconstruction at the decoder side. The high-quality reconstruction is controlled by transmitted parameters P which control the stereo-to-multi-channel upmix process. These parameters contain information describing, amongst others, the ratio of front versus surround signal which is present in the 2-channel down mix. Using such an approach, a decoder can control the amount of front versus surround signal in the upmix process. In other words, the parameters describe important properties of the spatial sound field which was present in the original multi-channel signal, but which is lost in the stereo mix due to the down-mix process.
The current invention relates to the possibility to use this parameterized spatial information to apply parameter-dependent, preferably invertible, post-processing on a 2-channel down-mix to enhance the downmix, such as the perceptual quality or spatial properties thereof.
An object of the present invention is to make post-processing of the down-mix possible after encoding, based upon the parameters as determined in the multi-channel encoder and still maintain the possibility of multi-channel decoding without influences of the post-processing.
This object is achieved by a method and a device for processing a stereo signal obtained from an encoder, which encoder encodes an N-channel (N>2) signal into left and a right signals and spatial parameters. The method comprises processing of said left and right channel signals in order to provide processed signals. The processing is controlled in dependence of said spatial parameters. The general idea is to use the spatial parameters obtained from an N-channel-to-stereo coder to control a certain post-processing algorithm. In this way, the stereo signal obtained from the encoder may be processed, for example for enhancing the spatial impression.
In an embodiment of the invention, the processing is controlled by a first parameter for each input channel, i.e. for each of the left and right signals, which first parameter is dependent on the spatial parameters. The first parameter may be a function of time and/or frequency. Thus, the system may have a variable amount of post-processing of which the actual amount of post-processing depends on the spatial parameters. The post-processing may be performed individually in different frequency bands. The encoder delivers independent spatial parameters describing the spatial image for a set of frequency bands. In that case, the first parameter may be frequency-dependent.
In another embodiment of the invention, the post-processing comprises adding a first, second and third signal in order to obtain said processed channel signals. The first signal includes the first input signal, i.e. the left or right signal, modified by a first transfer function, the second signal includes the first input signal modified by a second transfer function, and the third signal includes the second input signal, i.e. the right or left signal, modified by a third transfer function. The second transfer function may comprise said first parameter and a first filter function. The first transfer function may comprise a second parameter, whereby the sum of said first parameter and said second parameter can be unity. The third transfer function may comprise said first parameter of the second input signal and a second filter function.
The filter functions may be time-invariant.
In one specific embodiment, the signals may be described by the equation:
with α being a constant.
Using this representation, the filtering effect of the filter functions H1, H2, H3 and H4 is variable by varying the parameters wl, and wr. If both parameters have values equal to zero, the post-processed signals L0w, R0w are essentially equal to the stereo input signal pair L0, R0. On the other hand, if the parameters are +1, the post-processed stereo pair L0w, R0w, is fully processed by the filter functions H1, H2H3 and H4. This invention makes possible to control the actual amount of filtering, i.e., the value of the parameters wl, and wr by the spatial parameters P.
According to an embodiment, the filter functions and parameters are selected so that the transfer function matrix is invertible. This makes reconstruction of the original stereo signal possible.
In another aspect of the invention, it comprises a device for processing a stereo signal in accordance with the above mentioned methods, and an encoder apparatus comprising such a device.
In another aspect of the invention there is provided a method and a device for inverting the processing in accordance with the above mentioned methods, and a decoder apparatus comprising such an inverting device.
In yet another aspect of the invention there is provided an audio system comprising such an encoder apparatus and such a decoder apparatus.
Further objects, features and advantages of the invention will appear from the following detailed description of the invention with reference to embodiments thereof and with reference to the appended drawings, in which:
The encoded stereo signal pair L0 and R0 and decoder spatial information P, are transmitted to the user in a suitable way, such as by CD, DVD, VHS Hi-Fi, broadcast, laser disc, DBS, digital cable, Internet or any other transmission or distribution system, indicated by the circle line 4 in
However, due to the decreased number of playback signals, stereo signals are lacking spatial information compared to the N-channel signals or other properties that may be desired for certain situations. Thus, according to the present invention, there is provided a post-processor 5 which processes the stereo signal prior to the transmission/distribution to the receiver. The post-processing may be position-dependent “addition” of bass or reverberation, or removal of vocals (karaoke with vocals in center channel).
Other examples of post-processing are stereo-base-widening, which may be performed by making use of the knowledge of the composition of the original surround mix, such as front/back, since the contribution of individual input signals is known from the decoder information signals P. In principle, stereo widening can be applied already in the encoder, but this is generally not invertible, since only two signals are available in the decoder, instead of N, inversion is generally impossible. But besides stereo widening, also other post-processing techniques on the individual multi-channel contributions are possible.
According to the invention, the post-processed signals are transmitted to a receiver as indicated by the circle 6 in
The signal received may be used directly, for example if the receiver does not include a multi-channel decoder. This may be the case in a computer receiving the signal 6 over the Internet, or in a receiver having only two loudspeakers. Such received signal is perceived as a high quality signal, since it has improved spatial impression or other characteristics as determined in the processing thereof by the encoder and the post-processor.
If the signal should be used for decoding in a conventional N-channel decoder 3, it must first be inverse post-processed by an inverse post-processor 7, in order to reconstruct the original stereo signal pair L0 and R0 which together with the decoder information or spatial parameters P, produces an estimated N-channel signal. According to the invention, such reconstruction is possible of the multi-channel mix, which reconstruction is hardly affected by the post-processing. Also post-processing in the decoder is possible for stereo playback as a user-selectable feature, without the necessity to determine the multi-channel signal first. The inventive device for processing a stereo signal comprising left and right signals comprises the inverse post-processor 7. The decoder apparatus according to the present invention comprises the decoder 3 and the inverse post-processor 7.
Without post-processing the down-mix is comparable with a standard ITU down-mix. The inventive method, however, may improve the down-mix significantly.
The inventive method is able to determine the contribution in the down-mix of the original channels in the multi-channel mix with the help of the determined spatial parameters P in the encoder. In this way post-processing can be applied to specific channels of the multi-channel mix, for example stereo-base-widening of the rear channels, whilst the other channels are not affected. The post-processing does not affect the final multi-channel reconstruction if the post-processing is invertible. It can also be applied for an improved stereo playback without the necessity to reconstruct the multi-channel mix first.
This method differs from existing post-processing techniques in that it uses the knowledge of the original multi-channel mix, i.e. the determined spatial parameters P.
The encoder 2 operates in the following way:
Assume an N-channel audio signal as an input signal to the encoder 2, where z1[n], z2[n], . . . zN[n] describe the discrete time-domain waveforms of the N channels. These N signals are segmented using a common segmentation, preferably using overlapping analysis windows. Subsequently, each segment is converted to the frequency domain using a complex transform (e.g., FFT). However, complex filter-bank structures may also be appropriate to obtain time/frequency tiles. This process results in segmented, sub-band representations of the input signals which will be denoted by, Z1[k], Z2[k], . . . , ZN[k], with k denoting the frequency index.
From these N channels, 2 down-mix channels are created, being L0[k] and R0[k]. Each down-mix channel is a linear combination of the N input signals:
The parameters αi and βi are chosen such that the stereo signal consisting of L0[k] and R0[k] has a good stereo image. In case of a 5-channel input signal consisting of Lf, Rf, C, Ls, and Rs (for the left-front, right-front, center, left-surround, right-surround channels, respectively), a suitable downmix can be obtained according to:
L0[k]=L[k]+C[k]/√{square root over (2)}
R0[k]=R[k]+C[k]/√{square root over (2)}
The signals L and R can be obtained according to the equations:
L[k]=Lf[k]+Ls[k]/√{square root over (2)}
R[k]=Rf[k]+Rs[k]/√{square root over (2)}
Additionally, spatial parameters P are extracted to enable perceptual reconstruction of the signals Lf, Rf, C, Ls and Rs, from L0 and R0.
In an embodiment, the parameter set P includes inter-channel intensity differences (IIDs) and possibly inter-channel cross-correlation (ICCs) values between the signal pairs (Lf, Ls) and (Rf, Rs). The IID and ICC between the Lf, Ls pair are obtained according to the equations:
Here, (*) denotes the complex conjugation. For other signal pairs, similar equations can be used. Thus, the parameter IIDl describes the relative amount of energy between the left-front and left-surround channels and the parameter ICCl describes the amount of mutual correlation between the left-front and left-surround channels. These parameters essentially describe the perceptually relevant parameters between front and surround channels.
A parameterization of the amount of center signal which is present in L0, R0 can be obtained by estimating two prediction parameters c1, and c2. These two prediction parameters define a 2×3 matrix which controls the decoder upmix process from L0, R0 to L, C, and R;
An implementation of the upmix matrix M is given by:
For the example shown above, the parameter set P includes {c1, c2, IIDl, ICCl, IIDr, ICCr} for each time/frequency tile.
On the resulting stereo signal pair (L0, R0), post-processing can be applied in a way that it mainly affects the contribution of Zi[k], for example Ls, and Rs, in the stereo mix. In
This is shown in
The following equations hold for the post-processing parameters wl and wr:
wl=fl(IIDl,ICCl,c1,c2)
wr=fr(IIDr,ICCr,c1,c2)
The blocks H1, H2, H3 and H4 in
The resulting outputs are:
with α an arbitrary constant (e.g., +1).
If the filter functions H1, H2, H3 and H4 are chosen properly, the transfer function matrix H can be inverted. Moreover, to enable computation of the inverse matrix at the decoder side, the filter functions H1, H2, H3 and H4 and parameters wl and wr should be known at the decoder. This is possible since wl and wr can be calculated from the transmitted parameters. Thus, the original stereo signal L0, R0 will be available again which is necessary for decoding of the multi-channel mix.
Another possibility is to transmit the original stereo signal and apply the post-processing in the decoder to make improved stereo playback possible without the necessity to determine the multi-channel mix first.
Below, an embodiment of the post-processing is described in detail. However, the invention is not limited to the exact details but may be varied within the scope of invention as defined in the appended patent claims.
The post-processing parameters or weights wl, and wr are a function of the transmitted spatial parameters:
(w1,wr)=f(P)
The function f is designed in such a way that wl, increases if the signal L0 contains more energy from the left-surround signal compared to the left-front or center signals. In a similar way, wr increases with increasing relative energy of the right-surround signal present in R0. A convenient expression for wl, and wr is given by:
For the filter functions H1, H2, H3, and H4 the following exemplary functions are then chosen (in the z-domain):
H1(z)=H4(z)=0.8(1.0+0.2z−1+0.2z−2)
H2(z)=H3(z)=0.8(−1.0z−1−0.2z−2).
This invention can be integrated in a multi-channel audio encoder apparatus that creates a stereo-compatible down-mix. The general scheme of such a multi-channel parametric audio encoder which is enhanced by the post-processing scheme as described above can be outlined as follows:
A corresponding multi-channel decoder apparatus (i.e., a decoder with integrated post-processing inversion) can be outlined as follows:
Since the post-processing and inverse post-processing are performed in the frequency domain, the filter functions H1 to H4 are preferably converted or approximated in the frequency domain by simple (real-valued or complex) scale factors, which may be frequency dependent.
Those skilled in the art may understand that one or more processing stages as outlined above may be combined as a single processing stage.
Another application of the invention is to apply the post-processing on the stereo signal at the decoder-side only (i.e., without post-processing at the encoder side). Using this approach, the decoder can generate an enhanced stereo signal from a non-enhanced stereo signal.
Extra information can be provided in the bit-stream which signals whether or not the post-processing has been done and the parameter functions f1, f2 and which filter functions H1, H2, H3, and H4 have been used, which enables inverse post-processing.
A filter function may be described as a multiplication in the frequency domain. Since parameters are present for individual frequency bands, the invention may be implemented as simple, complex gains instead of filters, which are applied individually in different frequency bands. In this case, frequency bands of L0w, R0w are obtained by a simple (2×2) matrix multiplication from corresponding frequency bands from (L0,R0). The actual matrix entries are determined by the parameters and frequency domain representations of the filter functions H thus consisting of the time-invariant gains H and a time/frequency-variant parameter-controlled gains wl, and wr. Because the filters are scalars for each band, inversion is possible.
The post-processing in the encoder can be described by the following matrix equation:
This matrix equation is applied for each frequency band. The matrix H contains of all scalars. The use of scalars makes post-processing and the inverse post-processing relatively easy.
The parameters wl and wr, are scalars and functions of the parameter set P. These 2 parameters determine the amount of post-processing of the input channels.
The parameters H1 . . . H4 are complex filter functions.
The inversion of this process can also be done by a simple matrix multiplication per frequency band. The following equation is applied per frequency band:
The matrix H.sup.−1 contains only scalars. The elements of H−1, k1 . . . k4, are also functions of the parameter set P. When the functions in the matrix H, h11 . . . h22, and the parameters P are known in the decoder, then the post-processing can be inverted.
A block diagram of an inverse post-processor 3 which performs such inverse post-processing is illustrated in
This inversion is possible when the determinant of the matrix H is not equal to zero. The determinant of H is equal to:
det(H)=h11h22−h12h21=(1−wl)a(1−wr)a+(1−wl)awraH4+(1−wr)awlaH1+wlawra(H1H4−H2H3)
When suitable functions h11 . . . h22 are chosen, det(H) will be unequal zero, so the process is invertable.
It is mentioned that the expression “comprising” does not exclude other elements or steps and that “a” or “an” does not exclude a plurality of elements. Moreover, reference signs in the claims shall not be construed as limiting the scope of the claims.
Hereinabove, the invention has been described with reference to specific embodiments. However, the invention is not limited to the various embodiments described but may be amended and combined in different manners as is apparent to a skilled person reading the present specification.
Number | Date | Country | Kind |
---|---|---|---|
04101405 | Apr 2004 | EP | regional |
04103367 | Jul 2004 | EP | regional |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/IB2005/051065 | 3/30/2005 | WO | 00 | 10/2/2006 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2005/098826 | 10/20/2005 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
4095049 | Gerzon | Jun 1978 | A |
4236039 | Cooper | Nov 1980 | A |
5818941 | Embree | Oct 1998 | A |
6154545 | Kohut et al. | Nov 2000 | A |
6173061 | Norris et al. | Jan 2001 | B1 |
6205430 | Hui | Mar 2001 | B1 |
6697491 | Griesinger | Feb 2004 | B1 |
7447317 | Herre et al. | Nov 2008 | B2 |
20030236583 | Baumgarte et al. | Dec 2003 | A1 |
20050195981 | Faller et al. | Sep 2005 | A1 |
20070230710 | Van Loon et al. | Oct 2007 | A1 |
Number | Date | Country |
---|---|---|
9526083 | Sep 1995 | WO |
WO9526083 | Sep 1995 | WO |
9851126 | Nov 1998 | WO |
WO2004008805 | Jan 2004 | WO |
2005098821 | Oct 2005 | WO |
2005988241 | Oct 2005 | WO |
Entry |
---|
Faller et al: “Binaural-Cue Coding—Part II: Schemes and Applications”; IEEE Transactions on Speech and Audio Processing, vol. 11, No. 6, Nov. 2003, pp. 520-531. |
Schuijers et al: “Advances in Parametric Coding for High-Quality Audio”; AES Convention, Paper 5852, Mar. 2003, pp. 1-11. |
Hawksford: “Scalable Multichannel Coding With HRTF Enhancement for DVD and Virtual Sound Systems”: J. Audio Eng. Soc., vol. 50, No. 11, pp. 894-913. |
Warner Ten Kate: “Compatibility Matrixing of Multichannel Bit-Rate Reduced Audio Signals”; J. Audio Eng. Soc., vol. 44, No. 12, Dec. 1996, pp. 1104-1119. |
Warner Ten Kate: “A New Surround-Stereo-Surround Coding Technique”; J. Audio Eng. Soc., vol. 40, No. 5, 1992, pp. 376-383. |
Number | Date | Country | |
---|---|---|---|
20070183601 A1 | Aug 2007 | US |